1. Field of the Invention
The present invention relates to an image processing apparatus that discriminates attributes of each pixel and changes processing methods in accordance with a discrimination result. More particularly, the present invention relates to an image processing apparatus capable of precisely discriminating a color halftone dot region from other types of image regions.
2. Description of Related Art
In a digital copying machine or the like for reading an image by using a CCD sensor and decomposing the read image into image pixels, thereby carrying out image processing, an optimal image processing method differs depending on a character image, a photography image, and a halftone dot image. Because of this, in such a digital copying machine or the like, the image attribute is discriminated, whereby the image processing method is changed according to the discrimination result. Accordingly, precision is required when discriminating image attributes. Generally, an edge enhancement processing is applied to an image discriminated as a character image and a smoothing processing is applied to an image discriminated as a photographic image or a halftone dot image. Furthermore, It has been well-know that in case that an edge enhancement processing is to be applied to a character region in a halftone dot image, the edge enhancement processing is applied if halftone dots in the halftone dot image are color whereas a smoothing processing is applied if the halftone dots are black.
With the following reasons, an edge enhancement processing is applied to a character region in a halftone dot image in case that the halftone dots are color. That is, as shown in FIG. 24A, if black halftone dots are erroneously discriminated as a black character, proportional difference between black(K) and cyan(C), magenta(M), yellow(Y) becomes larger. Thereby, hue of a reproduction image is different from that of an original image and the reproduction image is likely to have image noises. On the other hand, as shown in FIG. 24B, even if color halftone dots are erroneously discriminated as a color character, proportion difference between black(K) and cyan(c), magenta(M), yellow(Y) does not change. Therefore, hue of a reproduction image does not differ compared with that of an original image. This is because image noises do not stand out in the reproduction image. Accordingly, discrimination of a character region in a black halftone dot region is not carried out but discrimination of a character region in a color halftone dot region is carried out. Therefore, edge enhancement processing is applied to a character region if the character region is in a color halftone dot region, thereby reproducing image with high quality.
However, in the above described conventional image processing apparatus, there has been a problem that if a color halftone dot image contains a black character in there, its color halftone dot region can hardly be discriminated as so precisely. That is, there has been a fear that color halftone dot regions around boundaries of a black character can hardly be discriminated precisely. This is because the number of color pixels existing in a predetermined region including a target pixel decreases around the boundary of the color halftone dot region and the black character region, whereby a count value of the color pixels approximates to a discrimination value. From this reason, there has been the fear that a color halftone dot region is erroneously discriminated as a black halftone dot region. If the color halftone dot region is erroneously discriminated as a black halftone dot region, a smoothing processing is also applied to a character region in the color halftone dot region. Therefore, even if it is a character in a color halftone dot image, an image of the character is not always reproduced precisely.